Re: [Networker] mminfo splitting data greater than 2 GB

Hi All,

The problem doesn't seem to be just related to the client and server
versions of Networker.

This has been tried on server/client version 6.0.x onwards till 7.1 and
the problem still
exists on almost all versions.However observed some interesting facts here.

If i query for let's say of an exchange database dump which is roughly
size of 60GB,
then there are no chunks generated and hence the output is OK as
below.While if the same
is done for any other unix filesystem e..g /backup of a Database server,
then the backup creates
chunks and the same is reported while querying.

The query below was done on a networker server running version 7.1
/opt/networker/bin/nsrd:
       $Revision: 92453-07 linker linker crt0.o B.11.30 010914 $
        Product:      NetWorker
        Release:      7.1.Build.230
        Build number: 230
        Build date:   Tue Sep 23 14:31:51 2003
        Build arch.:  hpux11n
        Build info:   DBG=0,OPT=

For Exchange server:
# mminfo -r
"client,totalsize(2),group,totalsize(2),level,sscreate(17),sscomp(17),pool,name,ssid,pssid"
-q "savetime>03/24/04 00:00:00,savetime<03/25/04 00:00:00" | grep -v
"index:" | grep -v bootstrap | grep -v "undefined" | sort -u | grep
exmbx01indi

exmbx01indi 81 GB exchg_group 81 GB incr 03/24/04 02:33:05 03/24/04
05:10:43 3Month X:\                         1264625161 0


For Unix Database server:
# mminfo -r
"client,totalsize(2),group,totalsize(2),level,sscreate(17),sscomp(17),pool,name,ssid,pssid"
-q "savetime>03/24/04 00:00:00,savetime<03/25/04 00:00:00" | grep -v
"index:" | grep -v bootstrap | grep -v "undefined" | sort -u | grep
doberman | grep backup | sort -k13

doberman 2000 MB RDBMS 2000 MB 9 03/24/04 23:37:55 03/24/04 23:49:11
3Month      /backup                        1717685883 0
doberman 2000 MB RDBMS 2000 MB 9 03/24/04 23:49:11 03/24/04 23:59:49
3Month      <1>/backup                     727830815  1717685883
doberman 2000 MB RDBMS 2000 MB 9 03/24/04 23:59:49 03/25/04 00:10:43
3Month      <2>/backup                     627168157  727830815
doberman 2000 MB RDBMS 2000 MB 9 03/25/04 00:10:43 03/25/04 00:23:33
3Month      <3>/backup                     274847275  627168157
doberman 2000 MB RDBMS 2000 MB 9 03/25/04 00:23:33 03/25/04 00:34:51
3Month      <4>/backup                     190961965  274847275
doberman 2000 MB RDBMS 2000 MB 9 03/25/04 00:34:51 03/25/04 00:47:52
3Month      <5>/backup                     140630995  190961965
doberman 2000 MB RDBMS 2000 MB 9 03/25/04 00:47:52 03/25/04 00:59:17
3Month      <6>/backup                     73522912   140630995
doberman 2000 MB RDBMS 2000 MB 9 03/25/04 00:59:17 03/25/04 01:10:43
3Month      <7>/backup                     4284604813 73522912
doberman 2000 MB RDBMS 2000 MB 9 03/25/04 01:10:43 03/25/04 01:21:48
3Month      <8>/backup                     4049724475 4284604813
doberman 1272 MB RDBMS 1272 MB 9 03/25/04 01:21:48 03/25/04 01:29:01
3Month      <9>/backup                     3965839060 4049724475

So what i can see from above is that the problem of continued savesets
still exists with newer versions.
I have included the ssid and pssid parameters for  reference as well.

Rgds,

Monit



Tim Mooney wrote:

In regard to: [Networker] mminfo splitting data greater than 2 GB, Nitin...:

I am using the following command:

mminfo -s $svr -r
"client,totalsize,group,totalsize,level,sscreate(17),sscomp(17),pool,nam
e" -q "savetime>$d2" | grep -v "index:" | grep -v bootstrap | grep -v
"undefined"| sort -u

The problem lies in that the mminfo splits up the save set of more than
2GB into chunks of 2GB and the rest, e.g. if a save set is of size 10GB
then it would give me 10 different outputs of 2GB sizes.


I think the blame is with the client (and maybe your server), not with
mminfo.  Older clients use something called continuation savesets, and
I think that's what you're running into.  If all your software is recent
(6.x or 7.x server, 6.x or later clients, very recent DBMS module), it
shouldn't be using continuation savesets anymore.

Where we see this is primarily with the NetWare client.  Continuation
savesets are a PITA because you can't use `name=whatever' in the mminfo
query, because it won't match any chunk after the first.

You don't report ssids with your report, and that may provide a valuable
clue.  As would the client and server versions.  As would the setting
for "SS cutoff size" in the Server->View->Details are of the GUI.

Tim

03/23/04 00:15:13          03/23/04 00:26:40             3Month
<1>/usr4


Yup, those look like continuation savesets, which means "old client".

Tim
--
Tim Mooney                              mooney AT dogbert.cc.ndsu.NoDak DOT edu
Information Technology Services         (701) 231-1076 (Voice)
Room 242-J6, IACC Building              (701) 231-8541 (Fax)
North Dakota State University, Fargo, ND 58105-5164

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=



--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=